Efficient conversion from Dependency Trees to Abstract Syntax Trees in Natural Language Processing

Abstract

This thesis is a contribution to the field of rule based Natural Language Processing. Within rule based NLP there are different grammar formalisms, each with their own strengths and weaknesses. This work presents several improvements to a tool for converting between two such grammar formalisms. First we improved the performance significantly. Secondly we improved the debugging facilities of the tool. Thirdly we made it possible to convert between trees that had previously been too structurally different from each other.